Economic theory and econometric method provide the appropriate constructs and measurement methods to examine alcohol treatment offset effects properly. The proposed research (1) begins with an application of economic and econometric methods through models to analyze both individual and provider behavior in the treatment of alcoholism, (2) provides a method for decomposing treatment costs into both direct and substitution effects, and applies the method to previous work, and (3) formulates and tests a rigorous econometric model of individual behavior concerning various types of alcoholism treatment. The first major task models treatment offset behavior in economic terms. This involves recognition of both consumer demand and provider behavior. The effort will distinguish among the various costs and benefits that accrue to alcoholism treatment. The second task concerns a framework with which to perform meta-analyses of existing studies to estimate such treatment offsets. This framework separates treatments into alcohol and non-alcohol related categories, and considers current and long term substitution and cost offsets. For example, current alcoholism treatment may decrease current or longer term utilization of other treatments. Impacts on costs, however, are related not only to utilization levels, but also to the relative costs of the comparative treatments. The third major task is the explicit estimation of alcohol treatment offset effects for individuals, using detailed data that enumerate cost, treatment, utilization, and insurance information for a large sample of those who are being treated for both alcohol-related and other disorders, to provide proper controls. Attempts will be made to distinguish the determinants of the decision to seek treatment, from the determinants of treatment levels. Policy-related scenarios will simulate the impacts of changes in costs of treatments, health insurance benefits for alcoholism treatment, and other parameters such as price, income, and sociodemographic risk factors. Data will come from a 12-year study of employees (and insured family members) of Deere & Co., collected by Holder (1 R01 AA06248). Study groups include (1) an experimental mental group of 1,000 treated alcoholics, (2) 2 control groups of untreated alcohol (each N=400). matched with the experimental group, (3) a comparison group of non-alcoholics (N=1,800).